Introduction

Accurate somatic mutation identification in NGS data is a critical step for a genomics approach since it can identify driver mutations and support clinical decision-making through target therapy. Different from germline variants, somatic mutations are the most common cause of cancer.

Methods

In this study, we described an effective pipeline to differentiate germline and somatic mutation using tumor-only Next-Generation Sequencing (NGS). In the first step, 2,320 total canine germline mutations were filtered out for somatic evaluation by comparison with previous publications and genomic databases. After that, mutations identified in more than five dogs with variant allele frequency (VAF) distribution centered at 0.5 (heterozygous germline) and 1.0 (homozygous germline) were filtered out (494 germline mutations). For mutations identified in less than five dogs, the protein sequence alignment between human and dog were compared. During this process 415 unclassifiable mutations were excluded.

Results

A total of 915 somatic mutations were identified using this filtering system. The VAF distribution of putative germline variants filtered out in this pipeline resembles known germline mutation, but not the TP53 and PIK3CA variants, which are somatic, indicating this is a valid method.

Conclusion

This is an effective and more affordable germline-somatic mutation discrimination methodology compared to traditional correction system that uses a matched normal sample from a healthy tissue of the same dog. Using our pipeline, we discovered canine mutations in 50 well-established oncogenes and tumor suppressors and compared them to those reported in human cancers.